What is the scientific problem that the nominee will address, and why is this important? What are the pioneering, and possibly high-risk, approaches that, if successful, might lead to groundbreaking results? Understanding brain function is a major outstanding question in biomedical science. Recently it has become clear that a major bottleneck in our progress in understanding of brain function is a gap, or rather an abyss, in our knowledge of the wiring diagram of the brain. This is a problem of fundamental importance because many neurological diseases arise from anomalies of the development of brain circuitry. Recent studies in schizophrenia, depression, bipolar disease, learning disabilities, autism, and X-linked mental retardations all point to developmental etiologies relating to aberrant circuit connections. Nevertheless, the normal connectivity of the brain is still a mystery and importantly, the methods for determining neuronal connectivity are severely limited. Failure to establish the requisite synaptic connectivity in brain circuits, or failure to refine topographic sensory or motor maps in response to experience, has devastating effects on circuit function. Nevertheless, how these brain circuits develop and change with experience has been difficult to ascertain due to limited methodologies. It is even more of a mystery how changes in the development of neuronal circuits can affect particular aspects of cognitive function and neurological health. A first step toward addressing these questions is to develop new technologies to determine neuronal connectivity in living animals. Over the next 5 years a major effort in my lab will be to develop a method to determine brain connectivity in vivo and to apply this method to address several fundamental questions pertaining to circuit formation and plasticity within the visual system. To know the underlying architecture of the brain, we must solve the [unreadable]who is connected to whom[unreadable] question. Because of its overarching importance, many labs are addressing this question. One recent approach has been to use serial section electron micrographs to identify and reconstruct neural networks. This method provides high resolution analysis, but it is labor intensive and only applicable to fixed tissue. Many investigators have concentrated on identifying reagents to cross synapses, such as wheat germ agglutinin [1], however one problem with this approach is that there is not a reliable way to amplify the signal in the downstream neurons. Without amplification, identification of downstream neurons poses a significant detection problem. Although some viruses may allow for amplification, I propose a different strategy based on the use of [unreadable]Trojan peptides[unreadable] and amplification through a modification of the Gal4/UAS system.